Convergent evolution of twintron-like configurations: One is never enough

Abstract

Introns inserted within introns are commonly referred to as twintrons, however the original definition for twintron implied that splicing of the external member of the twintron could only proceed upon splicing of the internal member. This review examines the various types of twintron-like arrangements that have been reported and assigns them to either nested or twintron categories that are subdivided further into subtypes based on differences of their mode of splicing. Twintron-like arrangements evolved independently by fortuitous events among different types of introns but once formed they offer opportunities for the evolution of new regulatory strategies and/or novel genetic elements.

Abbreviations

atpI

ATP synthase subunit a

cpn60

plastid-encoded chaperone gene

DMD

human dystrophin gene

fz

frizzled gene

groEL

the product of groEL gene is required for the proper folding of many proteins

HEG

homing endonuclease gene

IEP

intron encoded protein

kuz

kuzbanian gene

LATs

latency-associated transcripts

LC

lariate cap

ORF

open reading frame

osp

outspread gene

pros

Prospero gene

psbC

the gene encoding the CP-43 chlorophyll a-binding protein of Photosystem II

psbF

cytochrome b-559 gene

rnl

mitochondrial large subunit rRNA gene

rns

mitochondrial small subunit rRNA gene

rpoC2

gene encoding the β″-subunit of the plastid-encoded plastid RNA polymerase (PEP)

rps18

ribosomal protein S18 gene

rps3

ribosomal protein S3 gene

SSU

nuclear small subunit rRNA

Ubx

Ultrabithorax gene

ycf8

encodes a photosystem II polypeptide.

Introduction

Since the discovery that “intervening sequences” are removed during RNA processing from pre-mRNAs the complexity of intron splicing and the significance of introns toward gene regulation and for providing a platform for genes to express multiple products has been much appreciated. Today several categories of introns are recognized: introns requiring a spliceosomal machinery for removal, ribozyme type introns (“self-splicing” group I, II and III type introns), and the nuclear and archaeal tRNA introns that are processed by means of RNA endonucleases and RNA ligases. The term twintron, intron within an intron or introns occupying the same locus, was originally applied in situations where the internal intron had to splice first in order for the external intron to be reconstituted as a contiguous sequence and thus could become splicing competent. The possibility of a twintron configuration was first recognized for a complex group II intron in the psbF gene of Euglena gracilis by Michel et al.¹ but the term twintron was first applied for group II introns inserted within group II intron or group III introns.^2-7 Later other twintron variations were noted such as group I introns within group I introns,^8,9 group II introns within group I introns,⁹ lariat capping twin ribozyme introns,¹⁰ tRNA introns within tRNA introns ¹¹ and various combinations of spliceosomal introns embedded within spliceosomal introns.^12-16 In addition there are composite intron arrangements such as spliceosomal type introns inserted within a group I introns.⁸

With regards to twintrons composed of group I and group II introns (i.e. ribozymes) they may be the products of the mobility mechanisms of its components. Here the invading intron could have inserted into a critical component of the host intron that now prevents splicing of the external twintron component until the internal intron splices first. In other examples the invading intron may insert into segments not essential for splicing thus the external intron can splice independent of the internal intron.^9,17,18 In addition twintrons have been noted that are nested arrays of repeated insertions of an intron into another intron.^11,19 Finally, some group I or group II introns encode open reading frames (ORFs) and these ORFs can be invaded by introns. Although these internal introns may not perturb the splicing of the external intron, the invading intron disrupts the ORF and thus the internal intron has to be removed in order for the ORF to be expressed (i.e., mS1247 twintron; ¹⁸). Group I and II intron encoded proteins (IEP) are usually involved in intron mobility (reverse transcriptases, homing endonucleases), and/or in assisting intron RNA folding and splicing (maturases), and in some instances organellar introns can encode ribosomal proteins.^7,20-22

The functional significance for many of these ribozyme or spliceosomal twintron configurations is still being investigated. This review examines the various twintron-like (sensu stricto vs sensu lato) arrangements recorded so far and circumscribes them into better defined groupings. The various twintron-like arrangements along with examples of alternate splicing and trans-splicing events show that some genes have to be deciphered by various mechanisms at the RNA level.²³

Spliceosome twintron-like arrangements

Most eukaryotic protein coding genes are composed of exons and intervening sequences referred to as introns; both segments get transcribed but introns are removed by a multicomponent ribonucleoprotein complex called the spliceosome.²⁴ Two classes of spliceosomal introns are recognized: the U2-type and the rare U12 type.^25-27 The U2-type introns are characterized by consensus sequence features at the 5′ splice sites (AG/GURAGU), 3′ splice sites (YAG/G) and the branch site (CURA*CU; *= A branch point); and finally a pyrimidine-rich stretch is located between the branch point and the 3′ splice sites.^25,28 The U12 introns have poorly conserved 3′ splice sites but there are consensus 5′ splice sites (/RTATCCTTT) and branch sites (USCUUAA*CU; *= A branch point) sequences.²⁸ The U12-type introns lack the pyrimidine-rich region noted in the U2-type introns. One class of twintron observed in nuclear genomes of insects and recently vertebrates are composed of U12-types with their typical AU-AC termini nested within U2-type introns with GU-AG termini or versions where the U2-type introns are the external intron while the U12-type forms the internal component of the twintron arrangement (Fig. 1A;^14,29,30).

Figure 1.

Schematic representation of U2/U12 type splicesomal twintron-like arrangements. Up and downstream exons are represented by gray and black boxes respectively; internal and external introns are represented by blue and orange lines respectively; the intron to be spliced out is represented by a curved line. SD = sequence donor; SA = sequence acceptor. (A) U2/U12 type twintron (or nested intron); in some mixed spliceosomal twintrons either the internal intron is excised or the hosting intron but never both.^14,29 P = pyrimidine-rich stretch in U2 introns: located between the branch point and the 3′ splice sites. (B) Overlapping U2/U12 introns have been described,³⁰ here both the 5′ and 3′ splicing sites of one spliceosomal intron are positioned upstream of the splicing signals of the other spliceosomal intron. Note, although spliceosomal introns are represented as linear molecules they are released in a lariat configuration.

The classic example for this type of arrangement is the prospero (pros) gene in Drosophila, which encodes a transcription factor expressed in immature neuronal cells,^14,29,31 here a U2-type intron is embedded within a U12-type intron; however splicing of the U2-type intron precludes splicing of the U12-type. This intron arrangement allows for developmentally regulated alternative splicing whereby U2-type splicing dominates early during embryogenesis and U12-type splicing occurs later.²⁹ The decision as to which intron gets spliced appears to be regulated by the abundance of certain heterogeneous nuclear ribonucleoproteins during embryogenesis.²⁹

Evidence for U2/U12 type “twintrons” have been noted in other members of the Diptera ³² and a scan of the human genome identified potentially 18 such arrangements.³⁰ A study by Janice et al.³⁰ uncovered another variation of U2/U12 type “twintrons,” where the splice sites have shifted generating essentially overlapping introns. Here both the 5′ and 3′ splicing sites of one spliceosomal intron are positioned upstream of the splicing signals of the other spliceosomal intron (Fig. 1B). This “overlapping” arrangement could make splicing of both introns mutually exclusive as removal of one intron removes a splice junction for the other intron (Fig. 1B). The significance of these findings and the splicing pathways for the various U2/U12 twintron-like arrangements among the vertebrates are yet to be experimentally elucidated. If in some instances both nested introns belonging to different categories are removed during RNA processing these U2/U12 pre-mRNAs require 2 different large ribonucleoprotein complexes to operative simultaneous or in quick succession on the same transcript to scan the transcript.^26,30,33

Some twintron-like arrangement might have evolved to facilitate the removal of large introns. Large introns may have benefits such as housing nested genes and regulatory elements, but long introns can be problematic when it comes to being transcribed in a timely manner or during RNA processing.

Nested introns have been noted inside the large intron 7 (110 199 nt) within the human dystrophin gene (DMD). From the pre-mRNA 2 “internal/nested introns” are spliced before the remaining partially spliced intron is removed by the final splicing event that utilizes the 5′ and 3′ splice sites (Fig. 2A). The removal of “nested introns” might be a prerequisite to ensure that the “host introns” 5′ and 3′ splice sites come into closer proximity and thus ensuring the accurate removal of the intron.¹⁶ Removal of nested introns may also remove or reduce the possibility of the formation of RNA secondary structures that could interfere with the proper assemble of the spliceosome.

Figure 2.

Splicesomal twintron-like arrangements: (A) Nested introns within long introns. (B) Recursive splicing; RP = Ratchet points: intronic sequences (separated by “zero-length exons”) that are used as 5′ and 3′ splicing sites to remove large introns in smaller segments by recursive splicing. (C) Intrasplicing; Int. = Intraexon: a sequence with a dual role, in the first splicing reaction it behaves like an exon but in the second splicing reaction it is removed as part of an intron. (D) Fungal stwintron. The internal intron disrupts the 5' splicing site of the external intron (orange triangles); splicing of the internal intron allow for the reconstruction of the 5' splicing site of the external intron (orange square). Note, although spliceosomal introns are represented as linear molecules they are released in a lariat configuration.

Nested introns have also been found in “short” intervening sequences for example a 2 kb intron has been noted in the Herpes Simplex virus type 1 (HSV-1) latency-associated transcript (LAT).³⁴ LAT consists of 2 exons and one complex intron that can be processed in 2 mutually exclusive splicing pathways. In one, the entire 2 kb intron (2 kb) is removed; in the second pathway first a nested internal intron (0.5kb) is removed followed by the removal of the remaining 1.5 kb intron. In both scenarios the processed mRNA is identical; however the 0.5 kb RNA appears to have a shorter half-life compared to the 1.5 and 2.0 kb LAT intron RNAs. The function of LAT RNAs is not yet resolved but it has been speculated that they are regulatory in nature and maybe important in establishing latent infections in neurons of the sensory ganglia.³⁵ So the nested introns might provide a mechanism for generating different types of LAT RNAs in order for HSV-1 to maintain/or establish latent infections in certain neurons. Nested intron arrangements therefore may provide an avenue to generate a variety of RNA molecules that could be directly regulatory in function or be further processed into long or small non-coding regulatory RNAs.

Another strategy for removing long introns is exemplified by recursive (or ratchet) splicing first observed in large introns within Drosophila.^12,36,37 Here a large intron is removed from the pre-mRNA in multiple steps starting at the 5′ end of the large intron as segments that have abutting acceptor and donor sequences (Fig. 2B), these juxtaposed internal 3′ and 5′ splice sites are referred to as ratchet points (RPs). Ratchet point is used to indicate that the 3′ and 5′ splice sites define introns that are not “interrupted” by an exon ³⁷; sometimes this “exon” is referred to as zero-length exon.³⁸ It can be assumed that recursive splicing occurs co-transcriptionally thus a large intron gets shortened as transcription proceeds^36,37 until the intron is completely removed. This may have a stimulatory effect on the transcription complex and ensures that RNA polymerase II will transcribe long introns.^39,40

Recursive splicing has also been characterized in some long vertebrate genes (Homo sapiens) where due to the presence of cryptic splice sites or RP-sites long introns can be removed in a multi-step process.⁴¹ However, in long vertebrate genes recursive splicing also appears to involve what has been referred to as RP-exons, exons that may or may not be included in the final mRNA depending upon whether a functional 5′ splice site (RP-site) is assembled by a preceding splicing event. Failure of establishing a strong 5′ splice site will results in the inclusion of the RP-exon in the mRNA. Therefore recursive splicing can generate different mRNA isoforms depending on the generation of suitable RP-sites. Many RP-exons include early stop codons and thus their inclusion in the mRNA can decrease mRNA stability due to nonsense mediated decay, maybe RP-exons are a form of quality control marker to ensure proper recursively spliced transcripts are dominant over novel mRNA (RP-exon containing) isoforms.⁴¹

Some twintron-like configurations can be resolved by a mechanism referred to as intrasplicing, but this has only been observed in the vertebrate 4.1R- and 4.1B paralogs, which encode cytoskeletal adaptor proteins.^15,42 This “twintron-like arrangement” may have evolved to facilitate an alternate splicing event that allows for a distant (downstream) exon to be incorporated into the mRNA. Specifically it involves the coordination of alternative splicing of an intron located in the 5′ UTR region along with the utilization of an alternate promoter. In one alternative splicing scenario 2 nested introns are removed, first the downstream intron is spliced in a way that joins the central exon (intraexon) with another exon that is located >90 kb downstream. Joining of these 2 exons generates a splice acceptor site (3′ splice site) at their boundary and thus the second splicing event removes the first intron which now also includes the intraexon. So the intraexon plays a dual role, in the first splicing reaction it behaves like an exon but in the second splicing reaction it is removed as part of an intron, thus it is referred to as an intraexon. The removal of the downstream intron essentially allows for the removal of the upstream (i.e. first) intron along with the intraexon (Fig. 2C).

Superficially the recently described “exitrons” from Homo sapiens and Arabidopsis are a related phenomenon but here an internal component of an exon can be spliced out by utilizing cryptic splice sites, this allows for more alternative splicing options and more “plasticity” with regards to producing more variants for certain proteins.⁴³ In general nested introns, recursive splicing, and intrasplicing are in part fortuitous events where splice sites evolved to facility either the removal of large or complex introns in pieces or to increase the capacity for alternative splicing events.

Recently an arrangement referred to as spliceosomal twin introns (stwintrons) has been described from various fungal nuclear genomes.¹³ An internal intron disrupts the splice donor sequence of the external intron and therefore has to be removed in order to allow for the reconstitution of a functional splicing sequence in order to permit the external intron to be spliced out (Fig. 2D). This arrangement fits the strict definition of twintron where the internal intron has to be removed first in order for the second intron to be spliced out. At this point the origin or purpose/advantages of stwintrons is unknown; they may have evolved by chance via endogenous mechanisms. See Table 1 for the different categories of spliceosome twintron-like arrangements.

Table 1.

Different categories of spliceosome twintron-like arrangements.

Splicing mechanism	Organism	Host gene	Ref.
Twintron-like U2/U12 splicing	Drosophila melanogaster	pros	31
	Drosophila melanogaster	pros	14
	Drosophila melanogaster	pros	29
	Homo sapiens	18 genes(a)	30
Nested splicing	Homo sapiens	DMD	16
	Herpes Simplex virus	LAT	34
recursive splicing	Drosophila	Ubx	12
	Drosophila melanogaster	Ubx; kuz; osp; fz	74
	Drosophila	115 genes(b)	37
	Vertebrate (Homo sapiens)(c)	Genes expressed in the mammalian brain	41
intrasplicing	Vertebrate (Homo sapiens)(c)	4.1R	15,42
stwintron	Fusarium verticillioides Trichoderma reesei Botrytis cinerea	pih	13

^(a)See Supplementary Table 1 in Ref ³⁰ for details about the human genes harboring a twintron.

^(b)See Supplementary Table 1 in Ref ³⁷ for more details.

^(c)See Table 3 this paper and Ref. ^30,42

Ribozyme derived twintron and nested introns

Group I and group II introns are autocatalytic and can thus self-splice from transcripts although it has been recognized that in vivo splicing is enhanced or in some instances dependent on factors that can be intron encoded or are encoded by the host genome.^21,44-49 Both group I and group II introns show minimal primary sequence conservation, but they do have conserved secondary and tertiary structures. Group I and II introns can be distinguished by their splicing pathways and secondary/tertiary RNA folds.^1,50,51 Almost all group I intron RNAs contain paired regions referred to as P1 to P10, along with sequence segments (loops) that connect these helical regions. Group I introns can potentially self-splice by forming complex RNA structures were the P3 and P7 plus proximal P4, P5, P6 and P9 paired helical domains make up the catalytic core components and the P1 and P10 helices form the substrate domain and the later fold essentially brings the 5′ and 3′ splice sites into close proximity to each other.^52-54 Group I introns are classified into various categories based on conserved core regions and variations within the secondary and tertiary structures.^50,55-57 Group II intron RNA can be visualized as 6 stem-loop domains (referred to as domains I to VI) emerging from a central wheel (for a review, see references.^1,58-60 When ORFs are present they tend to be embedded within domain IV. Primary sequence conservation among group II introns is minimal except for domain V and the intron boundaries, with GUGYG and AY (Y = pyrimidines) defining the 5' and 3' ends, respectively.⁷ Group II introns are assigned to various categories in part based on structural features and the interactions of intron sequences with exon sequences in order to establish splicing competent folds.¹

Characterization of the chloroplast genome of E. gracilis revealed a plethora of group II and group III introns.⁶¹ Group III introns are essentially truncated group II introns that lost domains II to V ^{6, 62} and they typically range in size from 93 to 118 nt and splice in a similar manner as group II introns where the intron RNA is released as a lariat (for a review, see reference ⁶²). Twintrons were first described from Euglena cpDNAs,² and defined as genetic elements that splice such that the internal intron is removed first before the external intron can be spliced out (Fig. 3A). Among the euglenid plastids various twintron arrangements have been noted. Within the cytochrome b-559 (psbF locus) gene a group II intron has inserted within the domain V structural domain of a group II intron.² In the rps3 gene a group II intron inserted within a group III intron ⁶³ and within the rps18 gene (intron 2) a complex twintron consists of 4 group III introns where the outer intron is interrupted by an internal intron that contains 2 introns.⁵ In the psbC gene (intron 4) a group III intron is interrupted by a group III intron that contains an 1374 nt ORF.⁶⁴ Another unique twintron example was recorded within the E. gracilis chloroplast ycf8 gene where 2 internal group II introns are inserted in an external group II intron. Here in the predominant splicing pathway the 2 internal group II introns are spliced from subdomain ID of the external group II intron. What makes this twintron combination unique is that there is an alternative splicing pathway where the splicing of the first internal group II intron reconstitute a group III intron recruited from sequences of the external intron and the other internal intron.⁶ This example also provides evidence that group III introns can be assembled from segments of existing group II introns.

Figure 3.

Ribozyme twintron-like configurations: Group I, II and III introns are represented by red, green and purple lines respectively. 5'SS = 5' splicing site; 3'SS = 3' splicing site. Note, although group II or III intron RNAs are represented as linear molecules they are usually released in a lariat configuration. (A) Euglenid cpDNA type group III/group II twintrons; (B) mS917 twintron: fungal mtDNA group I type nested introns, rns = mitochondrial small subunit rRNA gene; HEG = homing endonuclease gene; (C) mS1247 twintron: fungal mtDNA group I/group II type nested introns; (D) mL2449 twintron: nested LAGLDADG type ORF encoding group I intron within an rps3 encoding group I intron; rnl = mitochondrial large subunit rRNA gene; (E) archaeal nested group II introns and possible splicing pathway; (F) group II nested introns recorded from Rhodophyta and cyanobacteria.

The various euglenid twintron arrangements are probably a result of the mobile nature of these types introns and the ability to maintain RNA folds that allow for splicing competent configurations that ensure their survival by minimizing their effects on the genes they have invaded.^4,65,66

The discovery of introns (“degenerated twintrons”) in cryptophyte plastids that resemble euglenid twintrons suggests that these elements can move laterally among unicellular eukaryotic algae that gained chloroplasts via secondary endosymbiosis.⁶⁷ However, the cryptophyte introns appear to be composed of external group II introns interrupted by a “degenerated group II intron-like sequence,” and the later does not appear to have splicing capacity resulting in an amalgamation with the outer intron.⁶⁷ So here 2 potential mobile introns merged to evolve into one splicing unit. A “twintron” – like element described by Meier et al. ⁶⁸ embedded within the cpn60 gene of Rhodomonas salina may represent a similar scenario as that described by Khan and Archibald ⁶⁷ for the groEL “degenerated twintrons” in species of Rhodomonas and Hemishelmis. The general consensus is that twintrons evolved by preexisting introns being invaded by other mobile introns. They are probably transitional in nature and can in some cases amalgamate to form new intron arrangements; i.e., from intron to twintron back to a single “complex” intron.

Various twintron-like configurations have recently been described from fungal mtDNAs (Figs. 3B,C&D; ^9,22). In Cryphonectria parasitica (Chestnut blight) and Ophiocordyceps tricentri within the mtDNA rns gene at position S917 a group ID intron has been invaded by another group ID intron that has an ORF embedded within its P2.1a loop. The internal group ID intron has inserted within the P1 loop of the external group ID intron. In Chaetomium thermophilum the mS1247 twintron is composed of an external group IC2 intron encoding an ORF that is interrupted by an ORF-less group II A1 intron.⁹ It was shown under in vitro conditions that the internal group IIA1 intron can self-splice and thus allows for the expression of a functional homing endonuclease protein encoded by the external group IC2 intron.¹⁸ A similar arrangement has been noted for the mL2449 intron in Grosmannia piceiperda, however here the external intron component is a group IA intron encoding a ribosomal protein (RPS3) and the internal intron is a group IC2 intron encoding a double motif LADLIDADG ORF (Fig. 3D). The internal intron is inserted within a sequence corresponding to the N-terminal segment of the external intron encoded rps3 ORF.²² RT-PCR showed that the internal intron is removed in vivo and this would allow for the expression of a functional RPS3 protein.²²

For the fungal mtDNA twintron-like elements described above there is no experimental or in silico evidence that the position of the internal introns could perturb the splicing capacity of the external introns so it might be best to view these types of arrangements as nested introns. Although, in the C. thermophilum mS1247 the internal group IIA1 intron interferes with the expression of the ORF embedded within the external intron this could be a regulatory strategy whereby expression of the intron encoded ORF is depend upon splicing of the internal intron.¹⁸ However, as it has been shown that in some cases IEPs can also serve as maturases.^47,48,49 Here the splicing of the internal intron could affect the splicing capacity/efficiency of the external or potentially both intron components.

Group II twintron-like arrangements have been recovered from Archaeal ^19,20 and Eubacterial genomes.^{7,17,21,69,70} In the Archaea some twintrons form large nested arrays by the repeated insertion of group II introns into one another (Fig. 3E). The target for internal introns appears to be a conserved sequence within the ORFs embedded in domain IV and thus introns keep inserting into the ORF located within this domain.¹⁹ Expansion and contraction of these nested introns could be due to unequal recombination events between nested arrays and/or due to homing of introns into domain IV ORF sequences.¹⁹ The first described bacterial twintron was described from the cyanobacterium Thermosynechococcus elongatus and here several group II (twin)introns where noted that contained insertions of a group II intron into domain IV of other group II introns (Fig. 3F; ^17,69). Another example of this twintron-like arrangement is found in Desulfitobacterium hafniense (GenBank accession no. NZ_AABB01000321), none of these twintrons have flanking exons that encode an ORF. Domain IV is a variable domain and not viewed to be a critical for splicing ^48,59; thus the prokaryotic “twintrons” probably should be referred to as nested introns unless experimental evidence shows that splicing of the internal intron is necessary for the splicing of the external intron.

An analogous arrangement to prokaryotic group II twintrons has been recorded for several strains of the unicellular red alga Porphyridium plastid genomes. Here group II introns were noted to be inserted within the domain IV of external introns ⁷¹; again no experimental data exists to show if splicing occurs in consecutive steps as in plastid twintrons previously described for some euglenids.^2,4

The most unique twintron-like arrangements are found within the nuclear rDNAs of the myxomycete Didymium iridis, the schizopyrenid amoeboflagellate Naegleria, and the Heterolobosea amoeba Allovahlkamphfia sp., which are distantly related protists.^10,72-76 Here a group I intron-like sequence is embedded within another group I intron. However, the internal intron has been co-opted into a lariat cap (LC) forming ribozyme that allows for an mRNA to be generated for the “intron” encoded and mobility promoting homing endonuclease protein. The mRNA is compatible with nuclear RNA processing and mRNA export to the cytoplasm for cytoplasmic translation (Fig. 4).^{10,54,72,77-80} The external intron is the group I “splicing” intron (GIR2) that allows for the maturation of the rRNA transcript.⁸¹ This arrangement is now referred to as a twin-ribozyme intron, as the internal ribozyme has lost its splicing capacity and evolved into the LC ribozyme. Overall these composite elements demonstrate that twintrons can evolve into complex multifunctional ribozymes. As a note of interest the twin ribozyme element from D. iridis also contains a small spliceosomal intron,⁸² so here ribozyme- and spliceosomal-type introns co-exist within one composite intervening sequence. See Table 2 for different categories of ribozyme-type twintron-like elements.

Figure 4.

Schematic representation of the twin ribozymes arrangement and its pre-mRNA processing in D. iridis protist. GIR1 = group I lariat capping ribozyme; GIR2 = group I splicing ribozyme; SI = spliceosomal intron; polyA = polyA addition signal required for the processing and addition of a poly A tail to the 3′ end of the mRNA. I-DirI = I-DirI homing endonuclease, BP = branching point; IPS = internal processing site involved in the 5′ lariat cap formation of the mRNA encoding I-DirI. The processing of the twin ribozyme transcript is described in Johansen 2002.⁷⁴

Table 2.

Ribozyme derived twintron and nested introns.

Twintron category	External intron	Internal intron	Organism	Host gene	Subcellular location	Ref.
GI/GI	GI(a)	GI(b)	Didymium iridis	SSU	Nucleus	8,77
	GI	GI(b)	Naegleria spp	SSU	Nucleus	72
	GI	GI(b)	Allovahlkampfia sp	SSU	Nucleus	10
	GID	GID	Cryphonectria parasitica Ophiocordyceps tricentri	rns	Mitochondria	9
GII/GII	GII	GII	Euglena gracilis	psbF	Chloroplast	2
	GII	GII(c)	Euglena gracilis	ycf8	Chloroplast	6
	GII	GII(d)	Pyrenomonas salina	cpn60	Chloroplast	68
	GII	GII(e)	Rhodomonas sp.	groEL	Chloroplast	67
	GII	GII(f)	Porphyridium purpureum	atpI rpoC2	Plastid	71
	GII	GII(f)	Thermosynechococcus elongates Desulfitobacterium hafniense	No ORF-encoding flanking sequences	Bacterial	17,69
	GII	GII	Methanosarcina acetivorans	No ORF-encoding flanking sequences	Archaea	19
GIII/GIII	GIII	GIII(g)	Euglena gracilis	rps18	Chloroplast	5
	GIII	GIII	Euglena gracilis	psbC	Chloroplast	64
GI/GII GII/GI	GIC2	GIIA1	Chaetomium thermophilum	rns	Mitochondria	9
	GIIA1	GIC2	Grosmannia piceiperda	rnl	Mitochondria	22
GIII/GII	GIII	GII	Euglena gracilis	rps3	Chloroplast	63

^(a)The ORF of the external group I intron also contains a small spliceosomal intron ⁸².

^(b)The internal intron has been co-opted into a lariat cap (LC) forming ribozyme.

^(c)The external intron is interrupted by 2 internal group II introns.

^(d)Degenerated group II intron-like sequence.

^(e)The internal intron has lost its splicing capability, resulting in an amalgamation with the outer intron.

^(f)Internal intron is inserted in D IV of the external intron.

^(g)The external group III intron is interrupted by an internal group III intron containing 2 additional group III introns.

Archaeal tRNA twintron-like configuration

There are twintron/nested intron examples for all types of introns (spliceosomal, group I, II and III) including Archaeal tRNA introns.^83,84 The tRNA introns are removed by RNA splicing endonucleases and the tRNA exons are ligated together by tRNA ligases.^84,85 So called multiple-intron-containing pretRNAs were described from Thermofilum pendens where introns are nested one inside the other (Fig. 5) and based on computational predictions the pretRNA folds into a conformation that would require the nested introns to be removed first in order for the external (canonical) intron to be excised.¹¹ This scenario fits the classic definition of a twintron as splicing occurs in a stepwise manner with internal introns being removed prior to the external intron.

Figure 5.

Archaeal tRNA twintron arrangement. The pretRNA contains multiple introns that are nested one inside the other (double/single nested). For maturation, the pretRNA folds into a conformation that allow efficient “splicing” of the external intron (i1) only after the splicing of the internal introns (i2 and i3). Sequential cleavage and ligation steps need to be carried out in order to remove the introns to produce the mature tRNA.

Proposed mechanisms for the formation of twintron-like elements

Twintron-like elements have been observed among viral, bacterial, archaeal and eukaryotic (nuclear, plastid and mitochondrial) genomes (See Table 3 for more details). It has been proposed that a variety of mechanisms can give rise to new introns such as mutations that can generate new splice sites leading to intronisation.^86-89 If the mutation-derived new splice site occurred within an intronic sequence that could eventually result in the formation of new introns within pre-existing introns (twintrons). The position of the newly formed intron (internal) within the pre-existing one will determine whether or not its splicing is essential for the splicing of the hosting (external) intron. Twintron-like configurations can evolve among mobile ribozymes as they are potential mobile elements that can be mobilized by their IEPs into ectopic sites or it has been postulated that group I and group II intron RNAs can potentially reverse splice into new sites by reverse splicing into RNA molecules.^90-94 The latter may be less specific compared to IEP mobility pathways as it only requires pairing of the internal guide sequence (typically 6 nucleotides within the P1 stem) for group I intron RNAs or exon/intron binding site (EBS/IBS) interactions for group II introns with the foreign RNAs. However, for both types of introns reverse splicing into RNA sequences would have to be followed by reverse transcription into cDNA and insertion of the cDNA into the genome by recombination.^94-97 For group II introns there is some evidence that the intron RNA can reverse splice into DNA sequences that resemble the intron's native homing sites (or IBS sequences).^98-101

Table 3:

Distribution of twintron-like arrangements among the three domains of life with reference to subcellular localization of each twintron category.

		Spliceosome twintrons					Ribozyme twintrons
Organism		U2/U12	Nested	Recursive	Intrasplicing	Stwintron	GI/GI	GII/GII	GIII/GIII	GI/GII	GII/GIII	tRNA twintrons
Homo sapiens(a) (b)	Metazoa	N	N	N	N
Drosophila sp.	Metazoa	N		N
Fusarium verticillioides	Fungi					N
Trichoderma reesei	Fungi					N
Ophiocordyceps tricentri	Fungi						M
Botrytis cinerea	Fungi					N
Chaetomium thermophilum	Fungi									M
Cryphonectria parasitica	Fungi						M
Grosmannia piceiperda	Fungi						M
Didymium iridis	Amoebozoa						N
Euglena gracilis	Euglenozoa							P	P		P
Allovahlkampfia sp	Heterolobosea						N
Naegleria spp.	Heterolobosea						N
Pyrenomonas salina	Cryptophyta							P
Rhodomonas Sp.	Cryptophyta							P
Porphyridium purpureum	Rhodophyta							P
Methanosarcina acetivorans	Archaea							A
Thermofilum pendens	Archaea											A
Desulfitobacterium hafniense	Bacteria							B
Thermosynechococcus elongates	Bacteria							B
Herpes Simplex virus type 1	Virus		V

(N) Nuclear; (M) Mitochondrial; (P) Plastidal; (A) Archaeal; (B) Bacterial; (V) Viral.

(a) In silico analysis suggests that U2/U12 type twintron arrangements are also present in other Vertebrates such as opossum, sloth, elephant, lizzards, chicken, lampreys, platypus, zebrafish, and others.³⁰

(b) Recursive splicing has recently been reported among human and other mammalian genes.⁴¹

Biotechnology

RNA mediated strategies for genome editing or silencing have received attention in recent years.^102-109 Group II intron based gene knock out systems referred to as targetrons have been successfully employed in many bacterial systems,¹¹⁰ and targetrons are also being evaluated for genome editing among eukaryotes.¹¹¹ Spliceosome-mediated RNA trans-splicing, which involves the recombination of 2 pre-mRNAs molecules, has been noted to have application for mRNA repair.¹¹² Trans-splicing group I introns also have the potential to be applied for RNA repair by replacing a mutated regions of an mRNA with a corrected sequence.^113-115 Trans-cleaving ribozymes also show promise as gene inactivation tools that can be used to remove disease associated host mRNAs or viral transcripts.^116-121

Twintrons offer new ribozyme scaffolds that can be engineered where the activity (or expression of ORFs) of the RNA can be regulated by the requirement of the splicing of an internal component, or the activity can be modified by alternative splicing of nested components allowing the ribozyme to assume alternate folds. With regards to ribozyme type twintron/nested configurations the potential exists that the ribozymes can be engineered with 2 different functional domains that perform different catalytic activities analogous to the twin-ribozyme introns.^114,122

Concluding remarks

The historical definition for twintrons is based on: 1. an intron is inserted within another intron and 2. the splicing of the internal intron is a prerequisite for the splicing of the external intron.² Most intron arrangements referred to as twintrons do not conform to both parts of this definition. In general twintron-like arrangements can be referred to as composite intervening sequences composed of 2 or more introns present at the same locus. For those intron arrangements where splicing of the internal intron is not essential for the splicing of the external member the term nested intron should be applied, while the term twintron should be applied to describe arrangements where splicing of the internal intron is a prerequisite for the splicing of the external intron (Fig. 6). In addition it would be advisable to indicate if the introns are spliceosomal, ribozyme based, or tRNA type introns (Fig. 6).

Figure 6.

Twintron-like arrangments can be divided into 2 major categories (twintrons and nested introns) based on the prerequisite of the splicing of the internal intron in order for the splicing of the external intron. Each category can be divided into several subdivisions based on the splicing mechanism: spliceosomal (U2 or U12 type splicing), ribozyme type (example shown is a group II intron embedded within a group I intron), or tRNA type splicing (requiring tRNA endonuclease and RNA ligases).

So U2/U12 (or U12/U2) type “twintrons” are best referred to as nested spliceosomal introns, some of the recently described fungal mtDNA “twintrons” (mS917, mS1247, mL2449) are best described as nested ribozyme introns, and tRNA nested introns that can splice independently of the external intron can be referred to as nested tRNA introns. Conversely elements that fit the original description of twintron can be referred to as spliceosomal twintrons (eg. stwintrons), ribozyme twintrons (eg. twintrons described from the cpDNA of euglenids) or tRNA twintrons.

Twintron-like configurations evolved multiple times among different intron categories and among spliceosomal introns such arrangements may be a platform for alternative splicing. Alternate mRNAs can be generated by splicing either one or the other intron component (or segment) from the pre-mRNAs, such as observed for the U2/U12 spliceosomal twintrons or in situations resembling intrasplicing. With regards to U2/U12 type introns it has been suggested that the slower splicing of the U12 intron may provide a way to slow down expression of the protein encoded by the mRNA.^33,123 The fungal stwintons may also provide a system where splicing of the internal intron can be a rate limiting step providing a fine tuning mechanisms for gene expression. Nested introns and precursive splicing may have evolved fortuitously by the formation and activation of cryptic splice sites, these events allow for the efficient removal of long introns and in some cases they may also offer a platform for alternative splicing events that can generate multiple mRNAs.

Ribozyme based twintrons may represent intermediate stages that favor the formation of new composite mobile elements, it can also provide an avenue for introns that lack ORFs to gain ORFs that can promote the mobility of the “twintron” like element. Ultimately as seen for spliceosomal twintron arrangements, internal introns could be regulatory elements, involved in controlling the mobility of the mobile intron and in some cases expression of the intron encoded proteins. For ribozyme type twintrons the combination of the ribozymes can lead to the formation of twin-ribozymes where the 2 components have evolved to perform different functions to allow for the expression of the internal “intron” ORF. “Intron shuffling” could be an important mechanism that can expand the repertoire of complex ribozymes, regulatory mechanisms, and mobile introns within many genomes.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors would like to acknowledge the helpful comment from 2 anonymous reviewers.

Funding

Research program of GH is supported through a Discovery Grant from the Natural Sciences and Engineering Research Council of Canada.